« Previous Article | Home | Next Article »
Doomed by Default: Tracing the Loss of Spaceshuttle Columbia
By: | Saturday, June 7th, 2008Doomed by Default
Tracing the loss of Columbia
by Kristine Mak Yu
Only the most ardent of space flight devotees tumbled out of bed to watch the flaring plume streak across the dawn sky that was the space shuttle Columbia reentering the Earth’s atmosphere on February 1, 2003.
 |
| A piece of the left wing. |
For Stanford students, it was the looming first round of Winter Quarter midterms that commanded their attention. Not the end of NASA’s 113th space shuttle mission, a low-priority science flight to complete a mishmash of experiments, including studies designed by schoolchildren about the effects of microgravity on bugs. The apathetic inattention that greeted Columbia upon its return indicated that, 17 years and 89 shuttle flights after the shocking Challenger explosion, space shuttle flight had once again become old hat to the American public. But shuttle flight means people blasting off to an orbital velocity of 25 times the speed of sound in less than 10 minutes in a machine built out of more than 2.5 million parts, drifting at the mercy of celestial mechanics dodging hurtling space debris, and then returning to ground in a massive braking sequence from orbital velocity to 220 mph which superheats the air to a couple thousand degrees Fahrenheit. It remains today an inherently risky rather than routine business. At 9:00:18 (ET) on the first day of February in 2003, a thunderous reminder of this risk came in the sonic booms as Columbia disintegrated and hurled debris across a few thousand square miles of Texas and Louisiana. All seven astronauts aboard perished.
“Columbia, Houston, UHF comm check.” A couple minutes after the explosion, the Mission Control team continued working futilely to re-contact Columbia. “Columbia, Houston, UHF comm check.” In the ionizing atmosphere of Columbia’s funeral pyre, Mission Control had been cut off from the shuttle 46 seconds before its main breakup and was unaware of the disaster. Minutes passed in a bewildered stasis until a cell phone rang: someone who had seen live TV coverage of the breakup had called a Mission Control team member and broken the news. In shock, NASA personnel mechanically set a contingency plan in motion that was the legacy of the Challenger disaster response. Controllers locked the door to Mission Control and archived all the mission data. A debris recovery team began operation. NASA Administrator Sean O’Keefe activated an accident investigation board. The forensics work had begun.
The appointed chair of the board, retired four-star admiral Hal Gehman, deftly took the helm, dubbing his team the Columbia Accident Investigation Board (CAIB). In response to criticism from Congress that the investigation board had the makings of an inside job, Gehman appointed in early March several new “university types,” including Douglas Osheroff, the current Stanford Physics Department chair and Nobel Laureate in physics. Osheroff’s appointment drew comparisons to that of Nobel Prize-winning physicist Richard Feynman to the Rogers Commission that investigated the Challenger explosion. During the Challenger investigation, Feynman famously plunked a deformed O-ring into ice water to demonstrate how cold temperatures robbed the material of its resilience and thus its ability to seal a joint. His elegantly simple demonstration focused the nation on the failure of the O-ring of the left solid fuel rocket booster as the direct cause of the Challenger explosion.
 |
| Investigators pieced together the charred remains of Columbia to determine the physical cause of the Columbia disaster. |
In an interview with The Stanford Report immediately after his appointment, though, Osheroff cautioned, “I am certainly no Richard Feynman. He was a brilliant physicist; I am a very good experimental physicist.” During Spring Quarter and the summer, Osheroff worked to uncover the physical cause of the accident, joining Gregory Kovacs, a Stanford professor in the Electrical Engineering department who was coordinating the CAIB debris and sensor analyses.
The physical cause for the disaster lay in the mangled innards that the failing Columbia had sprawled across seven states. The trail of debris recovered by a 250,000-person workforce after 1.5 million hours revealed that the left wing was the first part to be damaged and the most damaged. More clues came from the greatest find of the debris recovery: the miraculously intact flight recorder, which helped furnish a timeline for events leading to the shuttle breakup that pinpointed the site of a rupture in the left wing to heat-shielding Reinforced Carbon-Carbon (RCC) Panel 8. From the painstaking work of CAIB and NASA workers, all the lines of physical evidence inexorably converged to tell the same story: RCC Panel 8 on the leading edge of the left wing of the shuttle had been breached, allowing the superheated air upon re-entry to invade and incinerate Columbia from within. What caused the fateful breach of the shuttle’s thermal protection system was hardly a mystery. People inside and outside NASA had shaken accusatory fingers at foam from the infancy of the Columbia investigation. During liftoff, an external fuel tank that is half the height of Hoover Tower supplies supercooled liquid hydrogen and oxygen propellant to the main engines of the shuttle. A layer of foam is sprayed on the tank to keep the supercooled propellant from boiling and to prevent ice from forming on the tank surface that could break off and damage the orbiter. Just after Columbia’s launch, video cameras recorded a foam chunk falling off the external fuel tank and spraying debris in the vicinity of the left orbiter wing. Computational analysis estimated that a 1.67 pound chunk of foam from the left bipod ramp on the external tank had collided at 530 mph with RCC Panel 8, 81.9 seconds after launch.
Behind the crisp precision of decimal points lurked a muddled history of ignorance and denial in NASA’s dealings with foam. It exemplified the flawed NASA institutional culture–the default NASA modus operandi–that was the ultimate cause of the Columbia disaster. Eleven days after the shuttle disaster at a congressional hearing, O’Keefe downplayed the foam strike as a piffling dink: “The piece came off, dropped roughly 40 feet at a rate of something like 50 miles an hour, so it’s the functional equivalent, as one astronaut described to me, of a styrofoam cooler blowing off of a pickup truck ahead of you on a highway.”
 |
| On February 1, 2004, NASA lost the space shuttle Columbia and her crew, 17 years after the Challenger disaster. |
Even as the CAIB investigation dragged on into the summer with investigators checking off piece after piece of evidence implicating the foam strike, NASA was not fully convinced. Such unfounded denial can be attacked in two complementary ways: fundamental physical understanding that counter gut instinct and careful experiments to provide concrete data. A back-of-the-envelope calculation using basic physics produces a foam strike velocity two orders of magnitude higher than what O’Keefe suggested (see sidebar). The rebutting experiments to determine whether foam traveling roughly 500 mph could have caused damage to an RCC panel came courtesy of Scott Hubbard, a CAIB board member and the director of NASA Ames Research Center. “NASA absolutely had not a clue, I should say; they had never done these sorts of tests,” Osheroff pointed out in an interview. On a hot July day in San Antonio, a compressed nitrogen cannon fired a foam chunk at an RCC Panel 8 in a painstakingly designed recreation of the foam strike on Columbia. The foam struck and bashed a sixteen by seventeen inch hole in the panel. The NASA spectators, some with tears in their eyes, could no longer deny that a little piece of foam had meant the demise of Columbia and her crew.
 |
| A simulated foam strike tore a hole in the RCC panel during CAIB tests, confirming suspicions that loose foam downed the Columbia. |
NASA’s stubborn refusal to believe that foam strikes could seriously damage the shuttle turned out to be classic behavior. CAIB discovered that foam loss from the external tank had in fact occurred routinely without galvanizing a response since day one of space shuttle flight 25 years ago, on Columbia’s inaugural flight. Foam shedding has occurred in 65 of the 79 (82%) shuttle flights for which imagery is available. Once upon a time, foam strikes were considered a serious threat to the shuttle’s fragile thermal protection system. Engineers had stipulated that the delicate RCC panels were not to receive an impact greater than 0.006 ft-lbs, “which is probably a screwdriver being dropped from a distance of 6 inches” estimated Osheroff. As space shuttles kept going up, getting hit by foam, and coming down with no more than say, a hundred divots here and there, but oh, nothing irreparable, NASA came to accept foam strikes as an acceptable flight risk, as a maintenance rather than safety issue, without even understanding what the risk was. This disturbing pattern of institutional behavior was, like the foam strikes themselves, nothing new. In an interview on the Macneil/Lehrer Newshour in 1986 about his experiences investigating the Challenger disaster, Feynman described ridiculously similar behavior in NASA’s response to solid rocket booster defects:
I kind of imagined something like a child that runs in the road, and the parent is very upset and says, “It’s very dangerous!” The child comes back and says, “But nothing happened,” and he runs out in the road again, several times, and the parent keeps saying, “It’s dangerous!” Nothing happens. If the child’s view that nothing happened is a clue that nothing is going to happen, then there’s going to be an accident. You could hear brakes squealing a couple of times and that’s the leakage, then the gas going through the seals, and so forth. [. . . .] Sooner or later the child gets run over. Is it an accident? No, it’s not an accident.
Twenty years later, the brakes had squealed again in the form of especially large chunks of foam shedding from the left bipod ramp and in the egregious involvement of Columbia in five of the seven known foam shedding events from the left bipod ramp. NASA’s last sure chance to prevent the ensuing disaster had been two shuttle missions earlier in October 2002 during a launch of Atlantis. While NASA management had finally classified this foam strike as an action item, it had fatefully decided to wait just a couple more flights until resolving the foam problem.
NASA’s decisions regarding foam shedding rested on a shaky foundation. Osheroff discovered that NASA not only harbored an alarming lack of basic knowledge about the physical properties of foam it used but also had never properly analyzed the statistics of foam shedding.
 |
| Stanford professor Douglas Osheroff speaks at a Columbia Accident Investigation Board press conference. |
“I mean, they didn’t know how often bipod ramp foam had fallen off; they’d done no aero calculations to suggest what it might hit,” Osheroff huffed. A venerable tenet of NASA foam shedding mythology was that liquid propellant condensed in voids in the foam; after launch, aeroheating caused it to evaporate, expand, and then explode the foam. Osheroff took a look at the data from the last Atlantis launch and realized that the foam had fallen off long before aeroheating could occur. As NASA should have done long ago, Osheroff did experiments to see if pressure in voids could explode foam. On the loading dock of the Varian physics building at Stanford, Osheroff and his graduate student Jim Baumgardner took a piece of foam and glued it on a brass plate which had a hole in the center of it. Into this hole, they injected a liquid dye under gradually increasing pressure from a bike pump. Osheroff waited for a violent explosion to no avail. Puzzled, he bisected the foam and, observing the dye pattern, found in fact that there was a planar crack that extended all the way through the surface of the foam. In further kitchen sink experiments, Osheroff found similar results. So much for NASA’s exploding foam model: the pressure always escaped harmlessly through a crack. While he was working on his own fast and cheap, yet informative foam experiments, Osheroff found technicians at Marshall Space Flight Center laboriously setting up a series of brute-force experiments “that must have cost several million dollars” to simulate foam shedding from the left bipod ramp. They could not do it, as they should have known by examining the statistics that only 7 in the 79 foam shedding events in the past had been from the left bipod ramp.
In his appendix to the Rogers Commission’s final report, Feynman complained that shuttle engines were “designed and put together all at once” to be tested without prior testing of individual components. Two decades later, NASA technicians relied on such “top down” rather than “bottom up” methods like Osheroff’s simple experiments to elucidate foam shedding, as NASA management again took uninvestigated risks at the expense of safety, it seems that Challenger’s didactic legacy had been slowly whittled down to nothing more than safer solid rocket boosters and an official plan of action in the event of another disaster. The newfound attention to risk analysis and safety in shuttle missions in the aftermath of the Challenger disaster died away over the years.
 |
| The infamous bipod foam ramp. |
Additionally, NASA lost the probabilistic risk analysis reports on the shuttle thermal protection system that they commissioned in the early 1990s from Stanford Professor Elisabeth Pate-Cornell, now the head of the management science and engineering department. Together with then graduate student Paul Fischbeck, Pate-Cornell had clearly warned NASA that “they had to be careful with the insulation in the external tank and fix that soon.” But NASA only remembered the study after the Columbia disaster and “they did not remember where they had put it,” recounted Pate-Cornell. Similarly, the vaunted safety organization established after the Challenger disaster was, certainly by the time of the Columbia accident, regarded as largely a “silent safety organization.” Osheroff recalls:
“I got e-mail from [. . .] one of the NASA employees that oversaw the United Space Alliance [the safety organization] people as they were preparing shuttle stacks for launch. And he was told that his job was to prevent the USA people from writing up safety issues. So that’s exactly what he shouldn’t have been doing, of course. But the idea was that if USA wrote up something as unsafe, then in fact, it could delay the launch. So as a result of that, you’re compromising more or less on-time launch, and as a result, in fact, you are compromising the safety of the astronauts and the orbiter.”
The inescapable fact is that NASA has always been a political in addition to a scientific organization, beholden to the whims of the public and congressional funding. Delaying a launch has been the most consecrated taboo in NASA culture, and as long as schedule pressure drives NASA, more unnecessary deja vu disasters cannot be avoided. However what General Donald Kutyna, who held a role on the Challenger investigation board analogous to Gehman’s on CAIB, once said about launch schedules illustrates that schedule pressure will be difficult to eliminate:
“The shuttle schedules were very tight. It took a long time to process the flights, and there were only certain windows open in space-you could only launch at certain times, be it a certain time of the year or a certain time of day. If one shuttle launch was delayed for, say, mechanical reasons, it would delay all the shuttles in line following it. [. . . .] The other pressure on NASA to launch was the pressure from the press. The press had full coverage of the launches, and every time anyone dropped a screwdriver it would appear in the press: “Gee, we’ve goofed up, and we can’t fly.” Next, Congress gives you money according to how successful you are. They look at every mission you fly, and if the find something that is unsuccessful, that doesn’t test very well, then it affects your budget in the future.”
Osheroff happily reported that NASA is getting rid of the troublesome foam bipod ramp and adding a heater in its place to prevent ice buildup. Furthermore, NASA is finally starting to do ballistic experiments to properly determine how much impact an RCC panel can take. However, this is no indication that NASA will successfully fight off a regression to its default of staying on schedule at all costs. The support of the public, the attention of the press, and the financial backing of Congress will be crucial factors in NASA’s bid to reform. The current excitement about a manned mission to Mars may revive the swashbuckling glory of the Apollo days at best or allow a new set of impossible deadlines to ramp up the risk of another disaster if we do not heed the warnings of history.
Topics: CompSci, Engineering, and Design, Ethics, Environment, and Society, Physics, Volume 2, Issue 1 |
Comments
You must be logged in to post a comment.